Introduction: literature review
Research on interpersonal communication can and should contribute to the development of scientific findings. Yet, some scientific truths are created by imperfect humans using methods based on probabilistic inferences and puzzled with all sorts of potential for error. The problem of relying on interpersonal communication research, without some method of assessing errors, is that the results may not only produce inconsistent findings, but a chaotic theoretical approach to future research.
Allen (1997) suggested that many narrative or box-score reviews that try to make sense of interpersonal communication research could end up simply perpetuating errors and relying on chance. Additionally, the scholar suggested that attempts to explain inconsistencies in the interpersonal communication literature become more confused, especially when the number of studies becomes larger and larger. Allen (1998) explains that this happens because errors cannot be accounted for on the basis of methodological assumptions or some other type of examination of the investigations. Meta-analysis handles the issues of assessing the impact and the contribution to inter-study variability in outcome on the basis of random factors relating to sampling error.
Other interpersonal communication scholars such as Preiss and Allen (1995) argue that striving to formulate theories must also sift through all sorts of information riddled with various types of errors. They argue that a sophisticated examination, comparison, or classification does not provide a good basis for analysis without a statistical method for elimination of error such as meta-analysis. A telling example of the quandaries faced by these and other scholars summarizing large domains of complex research may be found in the similarity and attraction literature. Sixty years of accumulated literature resulted in competing theoretical camps and disagreements over germane processes and methodological approaches.
Dependent and independent variables
There are many approaches and possibilities for employing a meta-analytic review (Preiss & Allen, 1995). Occasionally, the results of a series of meta-analysis call into question an effort that is presumed to exist. For example, many scholars assumed that widespread differences in interpersonal communication were based on biological gender. That conclusion was called into question in communication by Canary and Hause (1993) and in the social sciences in general by Hyde and Plant (1995). The overall results led Allen (1998) to call for reconsidering both measurement and theoretical approaches. The examples of independent variables on interpersonal communication in his studies are self-esteem, power in language use, self-disclosure, communication process discussions, and interpersonal conflict.